Industrial Maturity of Machine Learning Solutions Within the Food Industry

Abstract

Ensuring food security is a crucial challenge becoming increasingly complex for society on a global level. Machine learning technology can help to overcome this challenge, however its successful deployment in practice is a mandatory prerequisite and currently achieved only to a limited extent. Therefore, this systematic literature review aims at determining the current state of industrial maturity of machine learning-based approaches in the context of food industry, evaluating their readiness for operational use and deployment. An initial framework for assessment of industrial technology readiness consisting of six technical and human- and process-related dimensions is developed. Existing solutions are categorized according to the addressed process step within the food value chain and the covered dimension of the maturity framework. As the findings demonstrate, the industrial maturity degree is mainly located in the lower to middle range. Regarding all considered dimensions and phases within the food value chain, however particularly regarding the dimensions integrability and usability as well as the phase packaging and logistics, huge progress is required to achieve an overall high or very high industrial maturity degree. Thus, this work highlights the importance of a holistic perspective realized e.g. by cooperation between research and industry in order to achieve application-ready machine learning models with high levels of industrial maturity.